A tape library device (LIB) that includes multiple recording media using magnetic tapes are widely used for use requiring a high-capacity storage for, for example, data backup for a host, such as a server. While magnetic tapes are advantageous in low costs and high storage capacity, the data access speed is lower than that of, for example, hard disk drives (HDD). For this reason, recently, hierarchical storage systems are used in each of which a disk array device including multiple HDDs is used as a cache mechanism (primary storage) and a tape library device having a much higher capacity is arranged at the backend (secondary storage). A hierarchical storage system can be referred to as a virtual tape library. A disk array device can be referred to as a tape volume cache (TVC). A tape medium can be referred to as a physical volume (PV).
Such a hierarchical storage system includes a hierarchy control server for controlling a disk array device and a tape library device. For example, when a data read request or a data write request is issued from a host, the hierarchy control server servers as a conventional tape library device for the host and writes or reads data by using the disk array device internally. The hierarchy control server regards a set of data acquired from a host as a logical volume (LV) and reads and writes the data to and from the disk array device. The hierarchy control server then saves, in a tape medium, the data that is written in the disk array device at the background not via the host. Specifically, the hierarchy control server saves, in a physical volume, i.e., a tape medium, a logical volume stored in the disk array device. The process of storing, in a physical volume, a logical volume in the disk array device can be referred to as “migration”.
In order to prevent the capacity of the hierarchical storage system from lowering due to the logical volumes that are stored in the disk array device, the hierarchical storage system does not perform updating and deletes the large-capacity logical volume having been subjected to migration.
For hierarchical storage systems, there is a technology for grouping used physical volumes in which logical volumes are already stored to generate a group called a stacked physical volume group (PVG) and managing the group. Furthermore, for hierarchical storage systems, there is a technology for grouping physical volumes not in use in which no data is stored to generate a group called a scratch PVG and managing the group.
Operations of migration performed by the hierarchical storage system will be described in detail here. The state where the logical volume to be read or written by the host from or in the disk array device is already stored in the disk array device can be referred to as a logical volume of an on-cache. The state where the logical volume to be read and written by the host from and to the disk array device is not stored in the disk array device can be referred to as a cache miss. When the logical volume to be read and write by the host is an on-cache, the hierarchy control server reads the logical volume from the disk array device and makes a response to the host. On the other hand, when the logical volume to be read and write by the host is a cache miss, the hierarchy control server inserts the physical volume to which the logical volume has migrated to the tape drive of the tape library and reads the logical volume. After transferring the read logical volume into the disk array device, the hierarchy control server makes a response to the host.
Furthermore, the hierarchical storage system updates performs a logical volume update in which a logical volume having data that is updated according to migration is stored in a physical volume. In order to omit time-consuming processing, such as generation management, an operation may be performed in which the old logical volume before update is regarded as invalid logical volume and the area of the invalid logical volume is regarded as an invalid area. When this processing is performed, each time the logical volume is updated, the invalid area in the physical volume increases and the usable area in the physical volume decreases. When it is not possible to store logical volumes in the physical volume anymore, a new physical volume not in use is used. For example, when a stacked PVG and a scratch PVG are used and if, in the stacked PVG, there is no physical volume in which a logical volume can be stored, a physical volume to be used is moved from the scratch PVG to the stacked PVG. Accordingly, the number of physical volumes not in use decreases eventually.
There is therefore a technology called reorganization as a conventional technology for increasing the number of physical volumes not in use. In reorganization, when the number of physical volumes not in use in a tape library device becomes small, logical volumes are transferred from the used physical volume to a physical volume not in use in order to increase the number of physical volumes not in use. Moving the stored logical volumes causes the used physical volume to turn into an physical volume not in use. When a stacked PVG and a scratch PVG are used, the physical volume not in use generated by reorganization is returned to the scratch PVG.
Note that there is a conventional technology called garbage collection in which data excluding invalid data is recorded in a new magnetic tape. Furthermore, there is a conventional technology in which a random medium, such as a magnetic disk, is used as a virtual sequential access recording medium and, when the random medium stores unneeded data, invalidation processing is performed. Furthermore, there is another conventional technology in which volumes are automatically reorganized in an optical library.
Patent Document 1: Japanese National Publication of International Patent Application No. 2010-522914
Patent Document 2: Japanese Laid-open Patent Publication No. 2008-146408
Patent Document 3: Japanese Laid-open Patent Publication No. 2010-282383
Patent Document 4: Japanese Laid-open Patent Publication No. 2002-297431
However, reorganization on multiple stacked PVGs may be executed simultaneously. In such a case, the reorganization is executed according to the order in which the hierarchy control server accepts process requests. When reorganization is executed, depending on the volume of data stored in a physical volume that belongs to a stacked PVG, a new physical volume from a scratch PVG may be secured. For this reason, when multiple sets of reorganization are executed simultaneously, it is assumed that it is not possible to secure a physical volume to be used for the reorganization process depending on the processing procedure. In such a case, no reorganization is performed and accordingly migration of a logical volume stored in a TVC to a physical volume becomes difficult and thus there is a risk that the redundancy of data in the hierarchical storage system is not realized.
Even with the conventional technology using garbage collection, it is not possible to prevent a shortage of physical volumes when multiple sets of reorganization are processed simultaneously, which makes it difficult to maintain redundancy of data in the hierarchical storage system. Furthermore, even with the conventional technology using a magnetic disk or an optical library, a similar problem occurs, which makes it difficult to maintain redundancy of data in the hierarchical storage system.